End-to-end internet speed analysis of mobile networks with mapreduce

Packet-based mobile networks are increasingly carrying internet data traffic for data intensive applications. Because of increased data usage, data traffic analysis have been a very essential way in network management and optimization for the mobile service providers. It is an important issue to measure user experiences in packet network, but the current systems cannot correctly capture user experiences. This is simply because; data traffic analysis is not end-to-end traffic trace. In the context of this study, the analysis system has been developed for mobile network. The test system runs speed tests with actual usage scenarios, trace communication packets and analysis download and upload. Test results as large data sets are processed and analyzed in parallel and distributed systems on cluster of computers using MapReduce programming model on Hadoop environment. Internet speed characteristics map is rendered with regional analyzed big test results. Additionally, the system reviews end-to-end network analysis and consequently, measure network speed quality too, with big data analysis, detects user experiences and problems without user complaints.